练手与防遗忘系列-numpy

越来越发现，numpy,pandas,scipy,matplotlib这类库的各种函数就像背英语单词一样，一直无法摆脱“背了忘、忘了背、背了再忘”的怪圈。为了使以后再使用的时候能够得心应手，不是那么费劲，特地记录一下练手的代码片段，以便以后可查。

变形与统计

# coding=utf8

import numpy as np

arr = np.array([1, 3, 8, 2, 4, 7])
print(arr)
print(arr.reshape([1, 6]))
print(arr.reshape([3, 2]))
print(arr.reshape([6, 1]))
print('-----------------------------------')
x = np.array([1, 2, 3])
print(x)
x1 = x.reshape([-1, 1])
print(x1)

x2 = x.reshape([1, -1])
print(x2)
print('-----------------------------------')
y = np.array([[1, 2, 3], [4, 5, 6]])
# 0维上的最小值 滑动0轴
print(y.min(axis=0))
# 1维上的最小值
print(y.min(axis=1))
print(y.mean(axis=1))
print(y.sum(axis=1))
print(y.prod(axis=1))
print(y.cumsum(axis=1))

运行结果：

[1 3 8 2 4 7]
[[1 3 8 2 4 7]]
[[1 3]
 [8 2]
 [4 7]]
[[1]
 [3]
 [8]
 [2]
 [4]
 [7]]
-----------------------------------
[1 2 3]
[[1]
 [2]
 [3]]
[[1 2 3]]
-----------------------------------
[1 2 3]
[1 4]
[2. 5.]
[ 6 15]
[  6 120]
[[ 1  3  6]
 [ 4  9 15]]

矩阵运算

# coding=utf8
# 矩阵运算

import numpy as np

arr = np.array([[1, 3, 8], [2, 4, 7]])
new = np.mat([[4, 5, 8], [3, 6, 10]])
print(arr + new)
print(arr - new)
print(arr / new)
print('-----------------------------------')
print(arr + 2)
print(arr - 4)
print(arr / 3)
print('-----------------------------------')
###################ndarray数乘和点乘#####################
arr2 = arr * 2
print(arr2)
print(type(arr))
print(type(arr2))
print(arr * arr2)
# 数量乘法
print(np.multiply(arr, arr2))
print(arr.dot(arr2.T))
print('-----------------------------------')
###################matrix数乘和点乘#####################
mat2 = new * 2
print(mat2)
# 数乘
print(np.multiply(new, mat2))
print(new * mat2.transpose())
print(new.dot(mat2.T))
print('----------------特殊矩阵-------------------')
print(np.zeros((2, 3)))
print(np.ones((2, 3)))
print(np.eye(3))
print(np.eye(3, 2))
print(np.identity(3))

运行结果：

[[ 5  8 16]
 [ 5 10 17]]
[[-3 -2  0]
 [-1 -2 -3]]
[[0.25       0.6        1.        ]
 [0.66666667 0.66666667 0.7       ]]
-----------------------------------
[[ 3  5 10]
 [ 4  6  9]]
[[-3 -1  4]
 [-2  0  3]]
[[0.33333333 1.         2.66666667]
 [0.66666667 1.33333333 2.33333333]]
-----------------------------------
[[ 2  6 16]
 [ 4  8 14]]
<class 'numpy.ndarray'>
<class 'numpy.ndarray'>
[[  2  18 128]
 [  8  32  98]]
[[  2  18 128]
 [  8  32  98]]
[[148 140]
 [140 138]]
-----------------------------------
[[ 8 10 16]
 [ 6 12 20]]
[[ 32  50 128]
 [ 18  72 200]]
[[210 244]
 [244 290]]
[[210 244]
 [244 290]]
----------------特殊矩阵-------------------
[[0. 0. 0.]
 [0. 0. 0.]]
[[1. 1. 1.]
 [1. 1. 1.]]
[[1. 0. 0.]
 [0. 1. 0.]
 [0. 0. 1.]]
[[1. 0.]
 [0. 1.]
 [0. 0.]]
[[1. 0. 0.]
 [0. 1. 0.]
 [0. 0. 1.]]

实用方法

# coding=utf8
# 实用方法

import numpy as np
import random

print(np.linspace(1, 9, 6))
data = np.random.rand(5)
print(data)
hist, bin_edges = np.histogram(data, [-1, 0, 0.3, 4])
print(hist, bin_edges)
print('-----------------------------------')
data = sorted(random.sample(range(100), 6))
print(data)
# 打印当前值应该插入的索引的位置
print(np.searchsorted(data, 8))
print(np.searchsorted(data, 25))
print('-----------------------------------')
# 随机生成具有正态分布的数组
print(np.random.rand(2, 3))
print('-----------------------------------')
a = np.ones((2, 2))
b = np.eye(2)
print(np.vstack((a, b)))
print(np.hstack((a, b)))

运行结果：

[1.  2.6 4.2 5.8 7.4 9. ]
[0.2564467  0.21826916 0.82746982 0.93122359 0.44223496]
[0 2 3] [-1.   0.   0.3  4. ]
-----------------------------------
[21, 30, 46, 50, 74, 83]
0
1
-----------------------------------
[[0.07296967 0.92940202 0.73958934]
 [0.25950902 0.950619   0.2525271 ]]
-----------------------------------
[[1. 1.]
 [1. 1.]
 [1. 0.]
 [0. 1.]]
[[1. 1. 1. 0.]
 [1. 1. 0. 1.]]